|
(11) | EP 1 408 549 B1 |
| (12) | EUROPEAN PATENT SPECIFICATION |
|
|
| (54) |
Process for manufacturing an array of cells including selection bipolar junction transistors and associated array of cells Herstellungsverfahren für Zellenanordnung mit bipolaren Auswahltransistoren und zugehörige Zellenanordnung Procédé de fabrication d'une matrice de cellules comprenant de transistors bipolaires du sélection et matrice de cellules associée |
|
|
||||||||||||||||||||||||||||
| Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention). |
[0001] The present invention relates to a process for manufacturing an array of cells including selection bipolar junction transistors. In particular, the invention refers to a memory array of a phase change memory (PCM) device, without being limited thereto.
[0002] As is known, phase change memory cells utilize a class of materials that have the unique property of being reversibly switcheable from one phase to another with measurable distinct resistivity. Specific materials that may be suitably used in phase change memory cells are alloys of elements of the VI group of the periodic table as Te or Se, also called chalcogenides or chalcogenic materials. Thus a thin film of chalcogenic material may be employed as a programmable resistor, switching between a high and a low resistance condition.
[0003] The use of chalcogenic storage elements has been already proposed to form a memory cell. To avoid disturbances caused by adjacent memory cells, the chalcogenic element is generally coupled with a selection element, generally a MOS transistor or a diode.
[0004] A possible organization of a PCM array is shown in Figure 1. The memory array 1 of Figure 1 comprises a plurality of memory cells 2, each including a storage element 3 of the phase change type and a selection element 4 formed here as a diode.
[0005] The memory cells 2 are arranged on rows and columns. In each memory cell 2, the storage element 3 has a first terminal connected to an own bit line BLn-1, BLn, BLn+1, ..., and a second terminal connected to an anode of the diode 4; the diode 4 has a cathode connected to an own word line WLn-1, WLn, WLn+1, ....
[0006] In order to address the storage element 3 belonging to a specific cell 2, for example the one connected to bit line BLn and to world line WLn, the bit line connected to the addressed cell (selected bit line BLn) is biased at a high voltage VOP and all the other (unselected) bit lines BLn-1, BLn+1, ... are grounded. Furthermore, the word line connected to the addressed cell (selected word line WLn) is grounded and all the other (unselected) word lines WLn-1, WLn+1, ... are biased at VCC, so that only the diode 4 connected to the selected word line and bit line is on.
[0007] CMOS compatible processes for manufacturing PCM have been already proposed, wherein the diodes are integrated in a P-type substrate wherein N-type regions are formed. The N-type regions, defining the cathode of the diodes, are contacted by a metal line and form wordlines of the array. The N-type regions accommodate P-type regions that define the anodes of the diodes and are connected to long stripes of chalchogenic material at small portions thereof defining the storage elements. The stripes of chalchogenic material extend perpendicular to the wordlines and define bitlines of the array.
[0008] Because of this structure, the diodes are associated with parasitic bipolar transistors having emitters formed by the diode anodes and connected to the bitlines (here, bitline BLn); bases formed by the diode cathodes and connected to the wordlines (here, wordline WLn); and collectors formed by the substrate. The equivalent electric diagram of a real cell is shown of Figure 2.
[0009] In practice, selection of a cell 2 is done by driving the bipolar transistor 4 in the active region and biasing the base-emitter junction in direct mode. Therefore, the actual current IB supplied by the base terminal is not equal to the current IE flowing through the emitter terminal, but is defined by the following relationship:
wherein βF is the current gain of the bipolar transistor.
[0010] The presence of these parasitic transistors causes some problems, mainly due to the high currents flowing in particular during a modify (set, reset) operation. Indeed, usually a write operation is performed on a number of cells on a single wordline (eight or sixteen cells, or even more) so that the total current flowing through the selected wordline and in the decoder drive pull-down transistor is the sum of the currents of the cells. This poses an upper limit to the number of cells that can be modified in a single operation, because the voltage drop across the selected wordline and the pull-down transistor becomes unacceptable.
[0011] Indeed, during a modify operation, a current in the range of 200 µA flows through the emitter terminal of each bipolar transistor 4. Since, as said, eight or sixteen cells connected to a same wordline are modified at the same time, the total current entering the emitters of the selected transistors is 1.6-3.2 mA.
[0012] Since in known memory arrays the bipolar transistor is not exploited, but instead it is considered a parasitic element, its design is not optimized, so that its current gain βF is much less than 1, the current flowing in the selected wordline is about the same as the total emitter current (1.6-3.2 mA, as above discussed); this current flows along the entire wordline and in the pull-down transistor of the row decoder, causing an unacceptable voltage drop.
[0013] A simple solution to this problem is that to divide the wordlines, introducing four or eight local drivers driving two bits as a maximum.
[0014] However, this solution increases the complexity of the memory device, the required area and thus the unitary costs of the memory device.
[0015] Another solution of the above problem resides in maximizing the current gain βF, so as to minimize the current flowing in the selected wordline and the pull-down transistor; thereby reducing the dimensions of the local driver or conversely the number of the local drivers to two or four, each local driver driving four bits.
[0016] However, the manufacture of a bipolar transistor as a selection element in a memory array with low current gain implies some problems. First of all, the process should be such as to avoid any possible shorting of the emitters and the base contacts in case of misalignments, even when the cell pitch is particularly tight.
[0017] Furthermore, a high current gain requires a very low base doping, leading to a high base resistance. Since the base is also the wordline of the array, both delay time and voltage drop along the wordline would be increased. Not least, a high current flowing into the substrate can give rise to injection problems that must be carefully handled. All these problems require an accurate optimization of the manufacturing process and a tradeoff between conflicting requirements.
[0018] US 2002/0079483 teaches an array of cells and the manufacturing method thereof according to the preamble of the main claims. Here, each cell comprises a diode stack, and the diode stacks are connected through first conductor lines which form, together with the epitaxial layer and the substrate, bipolar transistor. Furthermore, the array forms a non-planar structure, wherein the first conduction regions and part of the first conductor lines are formed in protrusions of the substrate, delimited on all four sides by two different insulations. Thus, this prior solution teaches a non-planar non-optimized structure, that shares the above discussed problems.
[0019] US 5,262,670 teaches a non-planar, complex structure, wherein each transistor is formed in a protruding portion of the substrate, and the wordlines are formed by polysilicon lines running laterally to the protrusions, isolated from the protrusions themselves (except for the basis regions) by oxide spacers. Each protrusion accommodates only one transistor and the control regions (base regions) extend only below the respective first conduction region and is not shared by different transistors. The manufacturing process is thus quite complex, and is not optimized.
[0020] Aim of this invention is to provide a manufacturing method allowing integration of a bipolar transistor in a cell array so as to overcome the above indicated drawbacks.
[0021] According to the present invention, there is provided a process for manufacturing an array of cells and an array of cells obtained thereby, as defined in claim 1 and, respectively, 11.
[0022] According to one aspect of the invention, the bipolar transistor operating as a selection element of an array of memory cells is formed in a semiconductor body including some distinct regions: a substrate of P-type; a subcollector region, of P-type, overlying the substrate; a collector region of P-type, overlying the subcollector region; a base region, of N-type, overlying the collector region; an emitter region, of P-type, housed in the base region; and a base contact. The emitter region and the base contact are contacted through contacts formed in a dielectric layer grown over the semiconductor body and are implanted after depositing the dielectric layer, or at least a lower portion thereof and opening the contacts, using an own mask. Thereby no misalignment of the emitter and base contact masks may cause short-circuits between the emitter and base contact regions.
[0023] According to another aspect of the invention, the doping of each of the regions forming the bipolar transistors is optimized as regards driving capability, leakage immunity, and low voltage requirement.
[0024] For the understanding of the present invention, preferred embodiments are now described, purely as a non-limitative examples, with reference to the enclosed drawings, wherein:
- Figure 1 illustrates a circuit diagram of a known array of phase change memory cells;
- Figure 2 shows an equivalent actual electric diagram of a memory cell of the array of Figure 1;
- Figure 3 shows the masks used for manufacturing a bipolar transistor according to a first embodiment of the invention;
- Figure 4 shows a cross-section of the first embodiment, taken along line IV-IV of Figure 3;
- Figure 5 shows a cross-section of the first embodiment, taken along line V-V of Figure 3;
- Figures 6-11 illustrate cross-sections taken along the same section plane of Figure 4 through a portion of a semiconductor wafer in subsequent manufacturing steps of a bipolar transistor of the cell array according to the invention;
- Figure 12 shows the doping profile for the bipolar transistor of Figure 4;
- Figure 13 shows the masks used for manufacturing a bipolar transistor according to a second embodiment of the invention;
- Figure 14 is a cross-section of the second embodiment, taken along line XIV-XIV of Figure 13;
- Figure 15 shows the masks used for manufacturing a bipolar transistor according to a third embodiment of the invention; and
- Figure 16 is a cross-section of the third embodiment, taken along line XVI-XVI of Figure 15.
[0025] According to the embodiment of Figures 3-5, a memory array is formed in a body 10 of semiconductor material including a heavily doped substrate 6 of P-type. A subcollector region 7, also of P-type and high doping level, extends above the substrate 6 in an epitaxial layer (not shown). A common collector region 11, of P-type and lower doping level than the subcollector region 7, extends on the substrate 6, also in the epitaxial layer. The epitaxial layer also houses a plurality of active area strips 12, of N-type, defining base regions. The active area strips 12 extend parallel to each other along a first direction (X-direction) and are electrically insulated from each other by field oxide regions 13 (Figure 5).
[0026] Each active area strip 12 accommodates a plurality of emitter regions 14, of P+-type, and a plurality of base contact regions 15, of N+-type, that are arranged alternately, that is each emitter region 14 is arranged between two base contact regions 15, and each base contact region 15 is arranged between two emitter regions 14. Thus, each pair of regions including an emitter region 14 and the adjacent base contact region 15 (for example, an emitter region 14 and the base contact region 15 arranged on right thereof), the active area strip 12 they are accommodated in, and the underlying common collector region 11 form a selection transistor 20 of PNP-type, corresponding to bipolar transistor 4 of Figure 1.
[0027] A dielectric region 21 extends on the body 10 and accommodates contacts, storage elements and interconnection lines. The dielectric region 21 is generally formed by more layers deposited subsequently to allow forming the various regions therein and may also include different materials.
[0028] First and second contacts 22, 23 extend in first and second openings 27a, 27b of the dielectric region 21. Preferably, the first and second contacts 22, 23 are of tungsten, covered on the vertical and bottom sides with a barrier material (for example, Ti/TiN), not shown for simplicity.
[0029] The first contacts 22 extend each from an emitter region 14 to a chalcogenic storage element 24 forming the storage element 3 of Figure 1. First metal lines 25, forming bit lines corresponding to bit lines BLn-1, BLn, BLn+1 of Figure 1, extend along a second direction (Y-direction), thus transversely to the active area strips 12. Each first metal line 25 is in contact with the chalcogenic storage elements 24 that are aligned in the Y direction, as visible from the cross-section of Figure 5. The first metal lines 25 are formed preferably in a first metal level.
[0030] The second contacts 23 are higher than the first contacts 22 and extend each from a base contact region 15 to second metal lines 26. The second metal lines 26, forming word lines corresponding to word lines WLn-1, WLn, WLn+1 of Figure 1, extend along the first direction (X-direction), thus parallel to the active area strips 12 and perpendicular to the first metal lines 25. Each second metal line 26 is in contact with the second contacts 23 that are aligned in the X direction, as visible from the cross-section of Figure 4. The second metal lines 26 are formed preferably in a second metal level.
[0031] Figure 3 shows some masks to be used for manufacturing the memory array of Figures 4 and 5. In particular, Figure 3 shows an active area mask 30, a contact mask 31 and an emitter mask 32.
[0033] As shown in Figure 6, the process starts from a semiconductor body or wafer 10 having a surface 10a and including a substrate 6 and an epitaxial layer 8, both of P-type, wherein the substrate 6 has a high doping (for example, higher than 1019 atoms/cm3) and the epitaxial layer has a low doping (for example, about 1015 atoms/cm3).
[0034] First of all, field oxide regions 13 (shown with dotted lines in Figure 6 and visible from the cross-section of Figure 9) are formed in a per se known manner in the epitaxial layer 8, using the active area mask 30 of Figure 3, and thus laterally delimiting the active area strips 12.
[0035] Then boron is implanted at high energy at a dose of 1013-1014 atoms/cm2 (Figure 7) so that, at the end of annealing, the subcollector region 7 extends at a depth comprised between 400 and 850 nm from the surface 10a of the semiconductor body 10 and has a doping level comprised between 1017 and 1019 atoms/cm3, with a peak at a depth of about 0.55 µm (see also Figure 12). Thereby, a maximum resistance of about 500 Ω and thus a maximum voltage drop of 100 mV (for a collector current of 200 µA during a reset pulse) are ensured.
[0036] After annealing, a P-well is implanted in the array part of the device, forming the common collector region 11, extending at a depth comprised between 200 and 400 nm from the surface 10a. Since the common collector region 11 should sustain a high current density and thus high level injection effects (e.g. Kirk effect) have to be carefully prevented, the doping level of the common collector region 11 is quite high, of the order of 1017-1018, with a peak of about 1018 at a depth of about 0.25 µm.
[0037] Thereafter, Figures 8, 9, the active area strips 12 are implanted with N-type doping agents, thus forming the base regions of bipolar transistors extending down from the surface 10a to about 170 nm. Conveniently, arsenic with an energy of 150-300 keV, preferably about 200 keV is implanted, so as to obtain a final doping level of between 5*1017 and 5*1018 atoms/cm3. Thereby it is ensured that the base regions 12 have a lower depth than the field oxide regions 13 and thus that there is an effective insulation between adjacent wordlines.
[0038] Furthermore, the base thickness and peak doping represent a good tradeoff between two conflicting requirements: on one hand, they should be low, so as to achieve a low base Gummel number (and thus a high gain) and to reduce to a minimum any leakage of the collector-base and emitter-base junctions on the other hand, they should be high to avoid the risk of pinch-off when a positive voltage is applied to the wordline (to deselect it). Moreover, the use of arsenic as a doping agent and the selected energy further insure a very sharp doping profile, thereby reducing the risk of pinch off while keeping the base-collector junction as shallow as possible (in particular it has to be shallower than the field oxide isolation).
[0039] Thereafter, Figure 10, the body 10 is covered by a first layer of insulating material, forming the bottom portion of the dielectric region 21, and contacts are opened using contact mask 31 of Figure 3, thereby forming the first openings 27a and the bottom portion of the second openings 27b. Then, a boron implant (P+-emitter implant) is made, using emitter mask 32, so as to form emitter regions 14 with a doping level of about 1019-1020 atoms/cm3 below the first contacts 22. The emitter implant is studied so as to keep the emitter regions 14 as shallow and as abrupt as possible (preferably, the emitter regions 14 have a depth of about 50 nm). Thus, BF2 is selected for the implant, thereby reducing channeling (through amorphisation) and ensuring a shallow implant without using a very low energy implant (11B at energy in the range of 1 KeV).
[0040] Thereafter, using an own mask not shown, that is the negative of the emitter mask 32, base contact regions 15 are implanted below the second contacts 23 using arsenic (or even phosphorus). In case, the base contact regions 15 may be doped before the emitter regions 14.
[0042] Then, Figure 11, the first openings 27a and the bottom part of the second openings 27b are filled with a barrier layer 34, e.g. Ti/TiN, and with tungsten 35.
[0043] The process continues with the necessary steps to form the memory elements, including forming the chalcogenic storage elements 24, the first metal lines 25, the second metal lines 26, the upper portion of the dielectric region 21 and the upper portion of the second contacts 23, e.g. as described in European patent application N. 01128461.9, to obtain the structure shown in Figures 4 and 5.
[0044] Alternatively, instead of the chalcogenic storage elements 24, other storage elements or other two- or three-terminal elements that are compatible with standard CMOS back-end processes may be formed.
[0045] According to a different embodiment, a doped region 28 of N type, having a higher doping level than the active area strips 12, is formed below each emitter region 14, as shown by broken lines in Figure 4. In this case, an N conductivity type determining agent, e.g. arsenic, is implanted using a dose close to that used for the active area strips 12 and employing the emitter mask 32, just after or just before the P+ emitter implant. Thereby, the base resistance and thus the emitter-to-base voltage drop are reduced, increasing also the immunity of the bipolar transistor against emitter-to-collector leakage and punch-through.
[0046] Figures 13, 14 show a different layout for a memory array having a selection element formed as a high-gain transistor, as above discussed. Here, in the X-direction, each emitter region 14 is separated by the adjacent emitter regions 14 by a base contact region 15 on one side (left in the drawings), and by a field oxide region 40 on the other side (right in the drawings). As shown, the active area mask 41 (Figure 13) has an grid-like pattern, and a field oxide region 40, having a grid-like shape, delimits a plurality of active regions 42 of rectangular shape. Each active region 42 accommodates only one base contact region 15 and two emitter regions 40, arranged on different sides of the base contact region 15 in the X-direction. Thus, each active region 42 accommodates two bipolar transistors 43 that share a same base contact region 15.
[0048] As visible from Figure 13, the shape of the active area mask 41 and that of the emitter mask 44 differ from the active area mask 30 and the emitter mask 32 of Figure 3; however, contact mask 31 is about the same as in Figure 3.
[0049] The manufacturing process, the doping levels and energies of the memory array of Figures 13 and 14 are the same described above with reference to Figures 3-12, with the only exception of the shape of the active area mask 41 and the emitter mask 44, as above outlined.
[0050] Also in the embodiment of Figures 13 and 14 a N-doped region 28 (not shown) may be provided below the emitter region 14, to reduce the base resistance and improve punch-through immunity.
[0051] With the embodiment of Figures 13 and 14, it is possible to save around 20% of silicon area with respect to the embodiment of Figures 3-5, even if the active area corners could introduce defectivity issues.
[0052] Figures 15 and 16 show a third embodiment, wherein adjacent emitter regions 14 are not separated by other formations (base contacts or insulating material), but their electrical separation is only ensured by the intrinsic base region (active area strips 12).
[0053] Specifically, here the active areas are formed as active area strips 12, analogously to the embodiment of Figures 3-5; but each base contact region 15 is formed every two emitter regions 14, analogously to the embodiment of Figures 13 and 14. Thus, each base contact region 15 forms two bipolar transistors 50 with the adjacent emitter regions 14.
[0054] The mask used to obtain the structure of Figure 16 are shown in Figure 15: as may be noted, the active area mask 30 is the same as in Figure 3 and the emitter mask 44 is the same as in Figure 13.
[0055] The manufacturing process of the memory array of Figures 15 and 16 is the same described above with reference to Figures 3-5, with the only exception of the shape of the emitter mask 44, as above outlined.
[0056] In the embodiment of Figures 15 and 16, it is possible to further reduce the area occupation, depending on the minimum distance attainable between two adjacent emitter regions 14; however, the presence of lateral parasitic PNP bipolar transistors (formed by two adjacent emitter regions 14 and the intermediate portion of the respective active area strip 12) renders this embodiment applicable only to solution including design measure to reduce the resulting leakage current.
[0057] According to a different embodiment, more than two emitter regions 14, e.g. four, eight, etc., may be arranged between consecutive base contact regions 15 without an oxide or base isolation between them. In this case, the area occupation is still reduced, but the current leakage problem is worsened and base resistance could become a limiting factor for the emitters located farther from the base contact.
[0059] The doping of the emitter regions 14 and base contact regions 15 through the openings formed in the dielectric layer 21 ensures self-alignement of these regions, and thus avoids the risk of short-circuit between them in case of mask misalignment.
[0060] The high doping of the subcollector 7 and its location directly on the substrate to connect the high doped common collector region 11 with the boron diffusing from the highly doped substrate 6 are very useful to reduce the voltage drop.
[0061] Also the high doping of the common collector region 11 ensures the latter to withstand a high current density, as required for a selection element operating as a proper bipolar transistor; thereby effects due to high level injection (such as Kirk effect) are prevented.
[0062] The given values for the base doping and the use of arsenic as a dopant ensure a good tradeoff among the various requirements, in particular as regards electrical insulation between adjacent wordlines, high gain, pinch-off and current leakages.
[0063] The implementation of a high gain bipolar transistor allows a reduction in the current flowing in the selected wordline, and thus a reduction in the area necessary to integrate the row decoders. The reduced wordline current reduces the voltage drop on the wordline, allowing a lower voltage operation as well as allows the implementation of longer wordlines, that is the possibility of selecting more bits with a single wordline, allowing for a more efficient memory area.
[0064] The reduced rowline current allows for less current density in minimum width wordlines, and thus an improved reliability.
[0065] Finally, it is clear that numerous variations and modifications may be made to the cell array as described and illustrated herein, all falling within the scope of the invention as defined in the attached claims.
providing a body (10) of semiconductor material of a first conductivity type;
forming, in said body, a common conduction region (11) of said first conductivity type;
forming, in said body, a plurality of insulating field oxide regions (13) delimiting a plurality of active area regions (12; 42) extending above said common conduction region;
forming, in said active area regions, control regions (12; 42) having a second conductivity type and a first doping level;
forming, on top of said body, an insulating layer (21) having first and second openings (27a, 27b);
implanting first portions of said active area regions through said first openings (27a) with a doping agent of said first conductivity type, thereby forming, in said active area regions, second conduction regions (14) of said first conductivity type, each said second conduction region (14) forming, together with said control region (12; 42) and said common conduction region (11), a selection bipolar transistor (20; 43; 50);
implanting second portions of said active area regions through said second openings (27b) with a doping agent of said second conductivity type, thereby forming contact regions (15) of said control regions having said second conductivity type and a second doping level, higher than said first doping level;
forming, on top of said body, a plurality of storage components (3), each storage component having a terminal connected to a respective second conduction region (14) and defining, together with said bipolar transistor, a cell (2) of said cell array,
wherein the step of providing a body comprises growing an epitaxial layer (8) of said first conductivity type and a third doping level on a substrate (6) of said first conductivity type and a fourth doping level higher than said third doping level,characterized in that said step of providing a body (10) comprises the steps of:
implanting, in said epitaxial layer (8)on top of said substrate (6), a subcollector region (7) of said first conductivity type and a fifth doping level, higher than said third doping level;
implanting, in said epitaxial layer on top of said subcollector region, said common conduction region (11) so that said common conduction region has a sixth doping level, higher than said third doping level and lower than said fifth doping level,
and in that said step of forming control regions (12; 42) comprises implanting doping agents so that each control region extends from the body surface (10a) to a lower depth than and is completely insulated from adjacent control regions by said field oxide insulating regions (13).
implanting arsenic in said active area regions.
a common conduction region (11), forming said first regions of said selection bipolar transistors;
a plurality of active area regions (12; 42) overlying said common region (11) and delimited by field oxide insulating regions (13);
said active area regions accommodating said control regions of said bipolar transistors, said second conduction regions (14) of said bipolar transistors as well as contact regions (15) of said control regions, said contact regions having said second conductivity type and a second doping level, higher than said first doping level, and being connected to biasing lines (26),
characterized in that said body (10) comprises a substrate (6) extending below said common conduction region (11) and having said first conductivity type and a third doping level, and a subcollector region (7) extending between said substrate (6) and said common conduction region (11), said subcollector region having said first conductivity type and a doping level higher than said common conduction region (11),and in that said each control region (12; 42) of said bipolar transistors extends from the body surface (10a) to a lower depth than said field oxide insulating regions (13) and is completely insulated from adjacent control regions by said field oxide insulating regions (13).
Bereitstellen eines Körpers (10) aus Halbleitermaterial eines ersten Leitfähigkeitstyps;
Bilden einer gemeinschaftlich leitenden Zone (11) des ersten Leitfähigkeitstyps in dem Körper;
Bilden einer Mehrzahl von isolierenden Feldoxidzonen (13), welche eine Mehrzahl von Aktivbereichszonen (12; 42) begrenzen, die sich oberhalb der gemeinschaftlich leitenden Zone erstrecken, in den Körper;
Bilden von Steuerzonen (12; 42) mit einem zweiten Leitfähigkeitstyp und einem Dotierungspegel in den Aktivbereichszonen;
Bilden einer Isolierschicht (21) mit ersten und zweiten Öffnungen (27a, 27b) auf der Oberseite des Körpers;
Implantieren erster Bereiche der Aktivbereichszonen durch die ersten Öffnungen (27a) hindurch mit einem Dotierstoff des ersten Leitfähigkeitstyps, wodurch in den Aktivbereichszonen zweite leitende Zonen (14) des ersten Leitfähigkeitstyps gebildet werden, wobei jede zweite leitende Zone (14) zusammen mit der Steuerzone (12; 42) und der gemeinschaftlich leitenden Zone (11) einen Selektionsbipolartransistor (20; 43; 50) bildet;
Implantieren zweiter Bereiche der Aktivbereichszonen durch die zweiten Öffnungen (27b) hindurch mit einem Dotierstoff des zweiten Leitfähigkeitstyps, wodurch Kontaktzonen (15) der ersten Steuerzonen mit dem zweiten Leitfähigkeitstyp und einem zweiten Dotierungspegel, der höher ist als der erste Dotierungspegel, gebildet werden;
Bilden einer Mehrzahl von Speicherkomponenten (3) auf der Oberseite des Körpers, wobei jede Speicherkomponente einen mit einem zugehörigen zweiten leitenden Bereich (14) verbundenen Anschluss aufweist und zusammen mit dem Bipolartransistor eine Zelle (2) der Zellenanordnung bildet;
wobei der Schritt des Bereitstellens eines Körpers das Züchten einer Epitaxieschicht (8) des ersten Leitfähigkeitstyps und eines dritten Dotierungspegels auf einem Substrat (6) des ersten Leitfähigkeitstyps und eines vierten Dotierungspegels, der höher ist als der dritte Dotierungspegel, aufweist,dadurch gekennzeichnet, dass der Schritt des Bereitstellens eines Körpers (10) folgende Schritte aufweist:
Implantieren einer Subkollektorzone (7) des ersten Leitfähigkeitstyps und eines fünften Dotierungspegels, der höher ist als der dritte Dotierungspegel, in der Epitaxieschicht (8) auf der Oberseite des Substrates (6);
Implantieren der gemeinschaftlich leitenden Zone (11) in der Epitaxieschicht auf der Oberseite der Subkollektorzone derart, dass die gemeinschaftlich leitende Zone einen sechsten Dotierungspegel hat, der höher als der dritte Dotierungspegel und niedriger als der fünfte Dotierungspegel ist,
und dass der Schritt des Bildens der Steuerzonen (12; 42) das Implantieren von Dotierstoffen umfasst, derart, dass sich jede Steuerzone von der Körperoberfläche (10a) zu einer geringeren Tiefe als die Feldoxidisolationszonen (13) erstreckt und von den benachbarten Steuerzonen durch die Feldoxidisolationszonen (13) vollständig isoliert ist.eine gemeinschaftlich leitende Zone (11), welche die ersten Zonen der Selektionsbipolartransistoren bildet;
eine Mehrzahl von Aktivbereichszonen (12; 42), welche über der gemeinschaftlichen Zone (11) liegen und durch Feldoxidisolationszonen (13) begrenzt sind;
wobei die Aktivbereichszonen die Steuerzonen der Bipolartransistoren, die zweiten leitenden Zonen (14) der Bipolartransistoren sowie Kontaktzonen (15) der Steuerzonen aufnehmen, wobei die Kontaktzonen den zweiten Leitfähigkeitstyp und einen zweiten Dotierungspegel, der höher ist als der erste Dotierungspegel, haben und mit Vorspannungsleitungen (26) verbunden sind,dadurch gekennzeichnet, dass der Körper (10) ein Substrat (6), das sich unterhalb der gemeinschaftlich leitenden Zone (11) erstreckt und den ersten Leitfähigkeitstyp und einen dritten Dotierungspegel hat, und eine sich zwischen dem Substrat (6) und der gemeinschaftlich leitenden Zone (11) erstreckende Subkollektorzone (7) aufweist, wobei die Subkollektorzone den ersten Leitfähigkeitstyp und einen Dotierungspegel, der höher ist als derjenige der gemeinschaftlich leitenden Zone (11) hat,
und dass jede Steuerzone (12; 42) der Bipolartransistoren sich von der Körperoberfläche (10a) bis zu einer Tiefe, die geringer ist als die der Feldoxidisolierzonen (13) erstreckt und von benachbarten Steuerzonen vollständig durch Feldoxidisolierzonen (13) isoliert ist.
fournir un corps (10) de matériau semi-conducteur d'un premier type de conductivité ;
former, dans ledit corps, une région de conduction commune (11) dudit premier type de conductivité ;
former, dans ledit corps, une pluralité de régions d'oxyde de champ isolantes (13) délimitant une pluralité de régions de surface active (12; 42) s'étendant au-dessus de ladite région de conduction commune ;
former, dans lesdites régions de surface active, des régions de commande (12; 42) ayant un deuxième type de conductivité et un premier niveau de dopant ;
former, par-dessus ledit corps, une couche isolante (21) ayant des premières et des deuxièmes ouvertures (27a, 27b) ;
implanter des premières parties desdites régions de surface active à travers lesdites premières ouvertures (27a) avec un agent dopant dudit premier type de conductivité, formant ainsi, dans lesdites régions de surface active, des deuxièmes régions de conduction (14) dudit premier type de conductivité, chacune desdites deuxièmes régions de conduction (14) formant, avec ladite région de commande (12; 42) et ladite région de conduction commune (11), un transistor bipolaire de sélection (20; 43; 50) ;
implanter des deuxièmes parties desdites régions de surface active à travers lesdites deuxièmes ouvertures (27b) avec un agent dopant dudit deuxième type de conductivité, formant ainsi des régions de contact (15) desdites régions de commande ayant ledit deuxième type de conductivité et un deuxième niveau de dopant, supérieur audit premier niveau de dopant ;
former, par-dessus ledit corps, une pluralité de composants de stockage (3), chaque composant de stockage ayant une borne connectée à une deuxième région de conduction respective (14) et définissant, avec ledit transistor bipolaire, une cellule (2) de ladite matrice de cellules,
dans lequel ladite étape consistant à fournir un corps comprend la croissance d'une couche épitaxiale (8) dudit premier type de conductivité et d'un troisième niveau de dopant sur un substrat (6) dudit premier type de conductivité et d'un quatrième niveau de dopant supérieur audit troisième niveau de dopant,caractérisé en ce que ladite étape consistant à fournir un corps (10) comprend les étapes consistant à :
implanter, dans ladite couche épitaxiale (8) par-dessus ledit substrat (6), une région sous collecteur (7) dudit premier type de conductivité et d'un cinquième niveau de dopant, supérieur audit troisième niveau de dopant ;
implanter, dans ladite couche épitaxiale par-dessus ladite région sous collecteur, ladite région de conduction commune (11) de telle sorte que ladite région de conduction commune ait un sixième niveau de dopant, supérieur audit troisième niveau de dopant et inférieur audit cinquième niveau de dopant,
et en ce que ladite étape consistant à former des régions de commande (12; 42) comprend implanter des agents dopants de telle sorte que chaque région de commande s'étende de la surface du corps (10a) jusqu'à une profondeur inférieure aux régions de commande adjacentes et complètement isolée des régions de commande adjacentes par lesdites régions d'oxyde de champ isolantes (13).
implanter de l'arsenic dans lesdites régions de surface active.
une région de conduction commune (11), formant lesdites premières régions desdits transistors bipolaires de sélection ;
une pluralité de régions de surface active (12; 42) placées sur ladite région commune (11) et délimitées par des régions isolantes d'oxyde de champ (13) ;
lesdites régions de surface active accueillant lesdites régions de commande desdits transistors bipolaires, lesdites deuxièmes régions de conduction (14) desdits transistors bipolaires ainsi que les régions de contact (15) desdites régions de commande, lesdites régions de contact ayant ledit deuxième type de conductivité et un deuxième niveau de dopant, supérieur audit premier niveau de dopant, et étant connectées aux lignes de polarisation (26),
caractérisé en ce que ledit corps (10) comprend un substrat (6) s'étendant sous ladite région de conduction commune (11) et ayant ledit premier type de conductivité et un troisième niveau de dopant, et une région sous collecteur (7) s'étendant entre ledit substrat (6) et ladite région de conduction commune (11), ladite région sous collecteur ayant ledit premier type de conductivité et un niveau de dopant supérieur à ladite région de conduction commune (11),et en ce que chacune desdites régions de commande (12; 42) desdits transistors bipolaires s'étend de la surface du corps (10a) jusqu'à une profondeur inférieure auxdites régions d'oxyde de champ isolantes (13) et est complètement isolée des régions de commande adjacentes par lesdites régions d'oxyde de champ isolantes (13).